In an Expert Focus article for WaterBriefing, Gary Adams, Operations Director for Water at Calisen, who previously led Northumbrian Water's £160m metering programme, takes a look at why smart meter installation counts are the easiest number to publish and the least useful for understanding whether AMP8 is working. Commissioning accuracy rates, workforce capability and accreditation levels alongside installation volumes are also needed to hit AMP8 target of 10.4 million smart meters by 2030.

Gary Adam: Since February, energy suppliers have had to pay customers £40 in automatic compensation under Ofgem’s Guaranteed Standards of Performance when a smart meter installation appointment fails for reasons within their control - an engineer arriving without the correct tools or equipment to undertake the job, for example. That's a regulatory price on avoidable failure, and water hasn't had that reckoning yet. AMP8’s target of 10.4 million smart meters by 2030 is the right ambition, but I’m not sure we’re being honest about what hitting it takes.

Funding gets a programme approved, but doesn’t get meters installed, commissioned and transmitting. That requires teams trained and capable of doing the job at pace. In August, Ofwat provisionally approved £3.4bn of additional investment. That investment is welcome – it just means more money going into the same constrained delivery capability. 

We talk a great deal about how many people we need. What we’re not asking is whether they can do the job to standard. The sector needs a 36% increase in its workforce to deliver AMP8's £100 billion investment, and in 2025 two-thirds of water engineers were already looking at other sectors. Recruitment is the easier part. I’ve seen fully qualified recruits struggle for months with the parts a qualification doesn’t teach: the confidence to complete an internal installation alone and the customer service skills that go with it. 

Workforce conversation stops too early in the delivery chain

Water isn’t recruiting for a single skillset. A technician swapping a meter outside a property, a plumber working inside it and a crew excavating the street for a boundary box are three different trades, each competing with energy, telecoms and rail for the same people. 

The workforce conversation stops too early in the delivery chain. We talk almost entirely about field capacity: technicians, plumbers, civil crews. But that's only the middle of it. Upstream, a meter never reaches a doorstep without planners scheduling the programme, sequencing routes and managing dependencies at scale. Downstream, it never counts as delivered without quantity surveyors and project managers managing cost, contract and commercial risk to completion. A chartered surveyor or experienced project manager is a multi-year qualification, not a fast-track apprenticeship, and neither is interchangeable with a technician's training. 

There’s a fourth requirement nobody recruited for. A meter technician’s job used to be mostly technical: fit the asset, make it work, leave. Now it’s a customer experience touchpoint, and for many households the meter install will be their only contact with their water company besides the bill. Under Ofwat’s C-MeX customer experience measure for 2025-30, two-thirds of a company’s score comes from surveys of customers who have been in contact, up from half in AMP7. AMP8 means more installs, so getting that interaction right is more important than ever.

CCW found complaints about smart meter installations up 48% in its 2024-25 report. That’s not surprising. Explaining the meter to an elderly customer, recognising vulnerability and adapting to it, de-escalating a doorstep situation are things technical accreditation doesn’t teach. Doorstep competence is only half of it. The risk is that we continue at pace while defects, incomplete commissioning and remediation work accumulate behind the installation curve. 

Under pressure to hit volume targets, we can treat commissioning as a formality. A correctly commissioned smart water meter requires five things to be true at once:

• the asset is linked to the right property record
• the installation evidence is captured and auditable
• the communications module has been validated against the live network rather than a bench test
• the first readings have been received and checked
• any exception triaged rather than left open

Miss any one of those and the meter exists on an installation report but not on a demand-management dashboard.

Meter installation is a fifteen-to-twenty-year problem - not a five-year one

This is a fifteen-to-twenty-year problem, not a five-year one, and energy proves it. Energy suppliers have been installing smart meters under a licence obligation since 2012 and the deadline has moved from 2019 to 2030. They must now pre-emptively replace around seven million communications hubs in the South and Central regions before 2G services end by 2033. That’s a fresh wave of installation work more than a decade after the rollout should have finished and years after the workforce built for it was assumed to be winding down. Not because the programme failed, but because the technology, network standards and skills to service them keep moving. 

Water will hit the same wall unless we build workforce renewal into the programme now, not retrofitting once the first meters start failing. That means treating re-accreditation as a continuous cycle rather than a one-off qualification, and building AI-assisted diagnostics into that training from the start. Those joining the sector over the next decade won't tolerate AI-enabled ways of working – they’ll expect them, the same way they expect a smartphone over a paper form.

The obvious objection is that you cannot pause a rollout while people become competent. The dates are set, the funding committed and the programme must move. Competence isn’t something you wait for. Either it gets built alongside volume or paid for afterwards in remedial visits, failed commissioning and customers who won’t let you through the door twice.

Workforce capability - training and re-accreditation should be taken as seriously as installation targets

None of this is unfixable, but it needs to be treated as a capability problem rather than a hiring one, with training and re-accreditation taken as seriously as installation targets. The reformed Growth and Skills Levy already allows short, modular apprenticeship units of 30 to 140 hours, aimed at exactly this kind of accreditation. The sector should be building commissioning competence into one now.

Some companies are already investing. Calisen has run its own academy for meter engineers since 2016, with refresher accreditation built in. Others are taking a longer view: at least one major water company now forecasts its skill needs five to ten years out and ties apprenticeship intake to that forecast, so each cohort answers an identified need rather than general recruitment. That’s a model the rest of the sector should examine.

The intake itself is worth rethinking. Rather than bidding harder in the same labour market as everyone else, one pool worth building a dedicated pipeline into is the roughly 13,000 service leavers who exit the armed forces each year.

There’s also a rehabilitation job to do with the existing workforce, not only new recruits. We have brilliant people who have done this for over 20 years and are now being asked for skills nobody expected of them. That’s our responsibility, not theirs. And how you train an 18-year-old isn’t how you train someone later in life.

Installation counts are the easiest number to publish and the least useful for understanding whether AMP8 is working. I’d rather see commissioning accuracy rates, workforce capability and accreditation levels alongside installation volumes. Those numbers tell you whether a programme will deliver outcomes.

A workforce that can commission at pace and right first time turns a meter estate this size from insight into outcomes. Without it, the sector will have bought a great deal of data and very little change. 

AMP8 will be judged on how many meters go in, and Ofwat already tracks how many stay connected. But it should also be judged on whether we built the capability, not just the headcount, to get that right first time and keep it there. That question is open and needs answering before delivery pressure makes it harder to fix.